Re: [libxml++] Add parse_memory(const unsigned char* contents) ?
- From: Murray Cumming <murrayc murrayc com>
- To: libxml++ <libxmlplusplus-general lists sourceforge net>
- Subject: Re: [libxml++] Add parse_memory(const unsigned char* contents) ?
- Date: Sat, 25 Dec 2004 17:01:22 +0100
I have done this now. Please do take a look, because API freeze is on
January 10th:
http://www.gnome.org/start/2.9/
I'm still not sure whether I should be using const unsigned char* or
const char*.
On Sat, 2004-12-18 at 16:28 +0100, Murray Cumming wrote:
> So, is "unsigned char*" the best type to use for a pointer to a byte
> array, that could text be of any encoding?
>
> Will it make sense for people to cast their UCS2 w_char* array to a
> unsigned char*? Aren't there little-endian/big-endian issues there?
>
> I'd like to avoid using a void*.
>
> On Thu, 2004-10-28 at 17:55 +0200, Murray Cumming wrote:
> > DomParser and SaxParser currently have a
> > void parse_memory(const Glib::ustring &contents)
> > method.
> >
> > libxml can parse documents in other encodings, but Glib::ustring is for
> > UTF-8. At a stretch it can contain some other encodings, but I don't think
> > it can contain UTF-16 or UCS2.
> >
> > So I think we should add
> > void parse_memory(const unsigned char* contents, const Glib::ustring&
> > encoding = Glib::ustring());
> >
> > When the encoding is not specified, then libxml will get the encoding from
> > the XML declaration, doing some encoding guessing to get that far.
> >
> > Unfortunately we can't add this as a virtual method in the base Parser
> > class, because that would break ABI, but we could add it as a non-virtual
> > method to each derived class.
>
--
Murray Cumming
murrayc murrayc com
www.murrayc.com
www.openismus.com
[
Date Prev][Date Next] [
Thread Prev][
Thread Next]
[
Thread Index]
[
Date Index]
[
Author Index]